Accuracy of Precipitation Forecasts (1st Place at JSM 2018 Data Expo)

I attended the Joint Statistical Meeting (JSM) 2018 in Vancouver, Canada. There I competed in the Sections on Statistical Computing and Graphics' Data Expo challenge, and my presentation on precipitation forecast accuracy won first prize amongst students. The presentation became a paper, and now it is set to be published in a special issue of Computational Statistics in 2021. PDF on page.

Executive summary: Rain forecasts consistently over-predict how often rain will fall. This is because the definition of a rainy day has a low bar: when 0.01 inches or more of precipitation falls. If this threshold is raised to 0.07 inches, accuracy improves dramatically by not defining days where it hardly rains as "rainy". Forecasts can be improved further by setting a unique threshold for each city.

Pursuing Good Leads with Random Forests (school project)

 

Executive Summary: This paper is a demonstration of how machine learning can be used to optimize pursuing leads in sales. In a bank’s phone marketing campaign, whether a sale was closed or not is predicted using a Random Forest based on customer data like history with the company, occupation and age. The final model correctly identifies a closed lead 69% of the time (True Positive Rate), and misidentifies a lead that doesn't close as one that does 17% of the time (False Positive Rate).

Analysis of a Likert survey in R

Which university departments have a more favorable view of Wikipedia usage?

Executive Summary: This is an analysis of a Likery survey on university faculty perceptions and practices of using Wikipedia as a teaching resource. In particular, departments will be clustered by how favorably they view Wikipedia usage according to 5 Likert scale questions. Categorical Principal Component Analysis was used to aggregate the questions, and then a Kruskal Wallis test and Dunn post hoc test were on the first principal component to cluster the departments. Two groups are distinguished from each other: Sciences, Engineering & Architecture and Arts & Humanities view Wikipedia more favorably than Health Sciences and Law & Politics.